How to refer a map column in a spark-sql query?

问题

scala> val map1 = spark.sql("select map('p1', 's1', 'p2', 's2')")

map1: org.apache.spark.sql.DataFrame = [map(p1, s1, p2, s2): map<string,string>]

scala> map1.show()

+--------------------+
| map(p1, s1, p2, s2)|
+--------------------+
|[p1 -> s1, p2 -> s2]|
+--------------------+
scala> spark.sql("select element_at(map1, 'p1')")

org.apache.spark.sql.AnalysisException: cannot resolve 'map1' given input columns: []; line 1 pos 18; 'Project [unresolvedalias('element_at('map1, p1), None)]

How can we reuse the dataframe map1 in second sql query?

回答1:

map1 is a dataframe with a single column of type map. This column has the name map(p1, s1, p2, s2). The dataframe can be queried for example with selectExpr:

map1.selectExpr("element_at(`map(p1, s1, p2, s2)`, 'p1')").show()

prints

+-----------------------------------+
|element_at(map(p1, s1, p2, s2), p1)|
+-----------------------------------+
|                                 s1|
+-----------------------------------+

Another option is to register the dataframe as temporary view and then use a sql query:

map1.createOrReplaceTempView("map1")
spark.sql("select element_at(`map(p1, s1, p2, s2)`, 'p1') from map1").show()

which prints the same result.

来源：https://stackoverflow.com/questions/64107171/how-to-refer-a-map-column-in-a-spark-sql-query